A high-resolution prediction system for birch pollen in Sapporo

Abstract

This study aims to develop a high-resolution forecast model for airborne birch pollen concentration in Sapporo. Using results that shows that the bulk of pollen grains was deposited near the source plant, we created a high-resolution birch tree density map from a field survey along public roads and footpaths and Street View on Google Maps. The map was linearly extrapolated to cover an area 10 km from the city centre, including forests in the south and southwest. A new atmospheric dispersion deposition model for birch pollen was developed to represent wind advection, turbulent mixing, and convection in the atmospheric boundary layer, and gravitational settling and scavenging with precipitation. Pollen emission was also modelled as a function of the birch tree density and surface air temperatures. The onset of the pollen season was based on the pollen observations by a sampler at Hokkaido Institute of Public Health in Sapporo, and the end of pollen season was assumed to be 40 days after its onset. Comparison with observations suggested that the hindcast simulation with the model reproduced the daily variations in the dry deposition well in most years from 2001 to 2011. The geographical distribution of birch pollen deposition was highly correlated with the birch tree density map, even on windy days.